Factors Influencing Conversion Kinetics of Triply-Bridged (μ-η¹:η¹-Peroxo)Diiron(III) Intermediates to Doubly-Bridged (μ-η¹:η¹-Peroxo)Diiron(III) Intermediates

Abstract

Several [Fe^II ₂(N-EtHPTB)(μ-O₂X)]²⁺ complexes (1•O₂X) have been synthesized, where N-EtHPTB is the anion of N,N,N′N′-tetrakis(2-benzimidazolylmethyl)-2-hydroxy-1,3-diaminopropane and O₂X is an oxyanion bridge. Crystal structures reveal five-coordinate (μ-alkoxo)diiron(II) cores. These diiron(II) complexes react with O₂ at low temperatures in CH₂Cl₂ (−90°) to form blue-green O₂ adducts that are best described as triply-bridged (μ-η¹:η¹-peroxo)diiron(III) species (2•O₂X). With one exception, all 2•O₂X intermediates convert irreversibly to doubly-bridged, blue (μ-η¹:η¹-peroxo)diiron(III) species (3•O₂X). Where possible, 2•O₂X and 3•O₂X intermediates were characterized using resonance Raman spectroscopy, showing respective ν_O-O values of ~850 and ~900 cm⁻¹. How the steric and electronic properties of O₂X affect conversion of 2•O₂X to 3•O₂X was examined. Stopped-flow analysis reveals that oxygenation kinetics of 1•O₂X are unaffected by the nature of O₂X and for the first time, the benzoate analog of 2•O₂X (2•O₂CPh) is observed.

Introduction

Nonheme diiron enzymes have attracted great interest because they perform a wide variety of reactions despite having very similar active sites.^1–8 For example, soluble methane monooxygenase inserts an oxygen atom into an alkane C-H bond, toluene and o-xylene monooxygenases, as well as phenol hydroxylase insert oxygen atoms into aromatic C-H bonds, the R2 subunit of class I ribonucleotide reductases extracts a hydrogen atom producing a stable radical and, as its name implies, stearoyl acyl carrier protein Δ⁹-desaturase dehydrogenates a fatty acid hydrocarbon chain. In addition to having similar active site structures, these enzymes all share one other property. The putative catalytic cycles of these enzymes all contain peroxide-bridged diiron(III) complexes formed upon reduction of molecular oxygen. In some cases, these intermediates are stable enough to be trapped and characterized.^9–17 While peroxo-diiron(III) species are endemic in these enzymes, their precise role in catalysis is poorly understood, even though they have been well studied. Synthetic attempts to study peroxo-diiron(III) intermediates have resulted in a variety of biomimetic complexes.^18–27 These compounds often form peroxide-bridged diiron(III) intermediates, especially when stabilized by carboxylate bridges. They are commonly reported as having a (μ-η¹:η¹-peroxo)diiron(III) motif.

In a previous publication,²⁸ we discussed how replacing the carboxylate bridges commonly employed in synthesis of these biomimetic complexes affected the behavior of the peroxo intermediates formed. Using the dinucleating ligand N-EtHPTB (anion of N,N,N′,N′-tetrakis(2-benzimidazolylmethyl)-2-hydroxy-1,3-diaminopropane) combined with two equivalents of iron(II) and one equivalent of either benzoate, diphenylphosphinate or dimethylarsinate,²⁹ we demonstrated that two different (μ-η¹:η¹-peroxo)diiron(III) species can be formed upon reduction of dioxygen. A doubly-bridged (μ-alkoxo)(μ-1,2-peroxo) intermediate was obtained from the benzoate-bridged diiron(II) precursor at −40°, in contrast to the triply-bridged (μ-alkoxo)(μ-1,2-peroxo)(μ-1,3-dimethylarsinato) intermediate produced by the dimethylarsinate-bridged diiron(II) precursor. The diphenylphosphinate-bridged diiron(II) precursor was unique in that it formed a triply-bridged peroxo intermediate as the initial kinetic product, which converted to a metastable doubly-bridged peroxo intermediate by a shift of the phosphinate from a bridging mode to a terminal monodentate binding mode before peroxo decomposition, analogous to the “carboxylate shift” notion described by Lippard twenty years ago.³⁰ We concluded that the nature of the oxyanion bridge (O₂X) strongly influenced the nature and stability of any (μ-η¹:η¹-peroxo)diiron(III) species formed, but we felt further work was required before the effects of bridge differences could be clearly understood. This work is intended to address that issue.

To this end, we focused on three effects: (i) those produced by O₂X O···O bite distance differences, (ii) those produced by O₂X electronic differences and (iii) those produced by O₂X steric differences. Examination required synthesis of several new diiron(II) complexes (1•O₂X) using N-EtHPTB and various O₂X ligands. We found that all of these new species reacted with O₂ in solution to produce blue-green (μ-η¹:η¹-peroxo)diiron(III) intermediates (2•O₂X) at −90°. With one exception, they then converted to deep-blue (μ-η¹:η¹-peroxo)diiron(III) species (3•O₂X) before decaying to yellow products (4•O₂X) at higher temperature. In most cases, addition of OPPh₃ to solutions of 2•O₂X led to conversion to purple-blue species (3′•O₂X) similar to 3•O₂X. In this paper, we report crystallographic details of three diiron(II) complexes, kinetic data on their reactions with dioxygen, and spectroscopic characterization of a variety of (μ-η¹:η¹-peroxo)diiron(III) intermediates. Implications of O₂X bite distances as well as electronic and steric differences are discussed.

Experimental Section

Materials and Syntheses

All reagents and solvents were purchased from commercial sources and were used as received, unless noted otherwise. The ligand N-EtHPTB was synthesized using a published procedure.³¹ Solvents were dried according to published procedures and distilled under Ar prior to use.³² The ¹⁸O₂ (97%) used in resonance Raman experiments was purchased from Cambridge Isotope Laboratories, Inc., Andover, MA. Preparation and handling of air sensitive materials were carried out under an inert atmosphere by using either standard Schlenk and vacuum line techniques or a glovebox. Elemental analyses were performed by Atlantic Microlab, Inc., Norcross, GA.

1•O₂CPh, 1•O₂PPh₂ and 1•O₂AsMe₂ were synthesized using published procedures.²⁸

1•O₂PMe₂

N-EtHPTB (157 mg, 0.217 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.19 mL, 1.4 mmol). Dimethylphosphinic acid (21.6 mg, 0.230 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN³³ (189 mg, 0.434 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (149 mg, 0.435 mmol) was added, resulting in immediate precipitation of a white powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced colorless crystals, some suitable for X-ray diffraction structural analysis. Yield: 153 mg (80%). Anal. for [Fe₂(N-EtHPTB)( O₂PMe₂)](BPh₄)(OTf) and calcd for C₇₀H₇₅BF₃Fe₂N₁₀O₆PS: C, 60.27; H, 5.42; N, 10.04%. Found: C, 59.82; H, 5.49; N, 10.38%.

1•O₂P(OPh)₂

N-EtHPTB (176 mg, 0.243 mmol) was dissolved in MeOH (~10 mL) along with Et3N (0.19 mL, 1.4 mmol). Diphenylphosphoric acid (67.7 mg, 0.271 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (222 mg, 0.509 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (173 mg, 0.506 mmol) was added, resulting in immediate precipitation of a white powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced milky crystals, some suitable for X-ray diffraction structural analysis. Yield: 331 mg (79%). Anal. for [Fe₂(N-EtHPTB)( O₂P(OPh)₂)](BPh₄)₂ and calcd for C₁₀₃H₉₉B₂Fe₂N₁₀O₅P: C, 71.87; H, 5.80; N, 8.14%. Found: C, 71.77; H, 5.94; N, 8.41%.

1•O₂CCPh₃

N-EtHPTB (208 mg, 0.288 mmol) was dissolved in MeOH (~10 mL) along with Et3N (0.19 mL, 1.4 mmol). Triphenylacetic acid (83.3 mg, 0.289 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (257 mg, 0.589 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (205 mg, 0.598 mmol) was added, resulting in immediate precipitation of a white powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced colorless crystals. Yield: 370 mg (73%). Anal. for [Fe₂(N-EtHPTB)( O₂CCPh₃)](BPh₄)₂ and calcd for C₁₁₁H₁₀₄B₂Fe₂N₁₀O₃: C, 75.78; H, 5.96; N, 7.96%. Found: C, 75.46; H, 5.95; N, 7.96%.

1•O₂CCMe₃

N-EtHPTB (153 mg, 0.212 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.19 mL, 1.4 mmol). Trimethylacetic acid (21.7 mg, 0.212 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (186 mg, 0.426 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (152 mg, 0.443 mmol) was added, resulting in immediate precipitation of a white powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced milky crystals. Yield: 130 mg (50%). Anal. for [Fe₂(N-EtHPTB)(O₂CCMe₃)](OTf)₂ and calcd for C₅₀H₅₈F₆Fe₂N₁₀O₉S₂: C, 48.71; H, 4.74; N, 11.36%. Found: C, 48.91; H, 4.77; N, 11.39%.

1•O₂CC₆H₂-3,4,5-(OMe)₃

N-EtHPTB (72.9 mg, 0.101 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.077 mL, 0.56 mmol). 3,4,5-Trimethoxybenzoic acid (21.4 mg, 0.101 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (94.2 mg, 0.216 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (87.2 mg, 0.255 mmol) was added, resulting in immediate precipitation of a pale yellow powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced yellow crystals, some suitable for X-ray diffraction structural analysis. Yield: 116 mg (74%). Anal. for [Fe₂(N-EtHPTB)(O₂CCH₂-3,4,5-(OMe)₃)](BPh₄)(OTf) and calcd for C₇₈H₈₀BF₃Fe₂N₁₀O₉S: C, 61.92; H, 5.33; N, 9.26%. Found: C, 62.14; H, 5.55; N, 8.94%.

1•O₂CC₆H₃-3,4-(OMe)₂

N-EtHPTB (100.0 mg, 0.138 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.097 mL, 0.69 mmol). 3,4-Dimethoxybenzoic acid (25.1 mg, 0.138 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (120.3 mg, 0.276 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (196.5 mg, 0.574 mmol) was added, resulting in immediate precipitation of a pale yellow powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced milky crystals. Yield: 162 mg (89%). Anal. for [Fe₂(N-EtHPTB)(O₂CC₆H₃-3,4-(OMe)₂)](OTf)₂ and calcd for C₅₄H₅₈F₆Fe₂N₁₀O₁₁S₂: C, 49.40; H, 4.45; N, 10.67%. Found: C, 49.65; H, 4.44; N, 10.39%.

1•O₂CC₆H₃-3,5-(OMe)₂

N-EtHPTB (146.1 mg, 0.202 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.142 mL, 1.02 mmol). 3,5-Dimethoxybenzoic acid (37.0 mg, 0.203 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (187.8 mg, 0.431 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (190.6 mg, 0.557 mmol) was added, resulting in immediate precipitation of a pale green-yellow powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced pale yellow crystals. Yield: 238.5 mg (80%). Anal. for [Fe₂(N-EtHPTB)(O₂CC₆H₃-3,5-(OMe)₂)](BPh₄)(OTf)₂ and calcd for C₇₇H₇₈BF₃Fe₂N₁₀O₈S: C, 62.36; H, 5.30; N, 9.44%. Found: C, 62.56; H, 5.11; N, 9.33%.

1•O₂CC₆H₄-4-OMe

N-EtHPTB (149.2 mg, 0.206 mmol) was dissolved in MeOH (~10 mL) along with Et₃N (0.144 mL, 1.03 mmol). 4-Methoxybenzoic acid (31.3 mg, 0.206 mmol) was added and allowed to dissolve. Fe(OTf)₂•2MeCN (189.2 mg, 0.433 mmol) was added, producing a yellow solution. After 5 minutes, NaBPh₄ (205.4 mg, 0.600 mmol) was added, resulting in immediate precipitation of a pale green-yellow powder. The solid was filtered and dried in vacuo. Recrystallization from MeCN and Et₂O produced milky crystals. Yield: 245.4 mg (93%). Anal. for [Fe₂(N-EtHPTB)(O₂CC₆H₄-4-OMe)](OTf)₂ and calcd for C₅₃H₅₆F₆Fe₂N₁₀O₁₀S₂: C, 49.62; H, 4.40; N, 10.92%. Found: C, 49.41; H, 4.51; N, 11.08%.

Physical Methods

UV-Vis spectra were recorded on a Hewlett-Packard 8453 diode array spectrophotometer (2-nm resolution) equipped with an Unisoku Scientific Instruments cryostat (Osaka, Japan). Resonance Raman spectra were collected on an ACTON AM-506M3 monochromator with a Princeton LN/CCD data collection system using a Spectra-Physics Model 2060 krypton laser. Low-temperature spectra of the peroxo intermediates in CH₂Cl₂ and MeCN were obtained at 77 K using a 135° backscattering geometry. Samples were frozen onto a gold-plated copper cold finger in thermal contact with a Dewar flask containing liquid nitrogen. Raman frequencies were referenced to the features of indene. Slits were set for a band-pass of 4 cm⁻¹ for all spectra.

Time-resolved spectra of rapid oxygenation reactions were acquired with a Hi-Tech Scientific (Salisbury, Wiltshire, UK) SF-43 multi-mixing anaerobic cryogenic stopped-flow instrument combined with either a monochromator (low intensity light irradiation of the sample) or a diode array rapid scanning unit (strong UV-Vis irradiation of the sample), or with a TgK Scientific (formerly HiTech Scientific, Salisbury, Wiltshire, UK) SF-61DX2 cryogenic stopped-flow system equipped with a J&M Diode array (Spectralytics). All manipulations with diiron(II) complexes and their solutions were done using an argon atmosphere glove-box, air-tight syringes, and the anaerobic stopped-flow instrument to avoid contamination with air. Saturated solutions of O₂ in CH₂Cl₂ and CH₃CN were prepared by bubbling the dry O₂ gas for 20 min in a septum-closed cylinder with the solvent at a constant temperature (20 or 25 °C). The solubility of O₂ was accepted to be 5.8 mM in dichloromethane at 20° and 8.1 mM in acetonitrile at 25°.³⁴ Solutions of O₂ with smaller concentrations were prepared by diluting the saturated O₂ solution with argon-saturated solvent using graduated gas-tight syringes equipped with three-way valves. For the kinetic experiments, dioxygen was always taken in large excess so that its concentration did not change significantly during the reaction with 1•O₂X. The solutions of 1•O₂X and O₂ were cooled to a preset temperature (±0.1°) in the stopped-flow instrument before mixing. Data analysis was performed with the IS-2 Rapid Kinetics Software (Hi-Tech Scientific) for kinetic traces at a single wavelength.

X-ray Crystallography

X-ray diffraction data were collected on a Bruker SMART platform CCD diffractometer at 173(2) K.³⁵ Preliminary sets of cell constants were calculated from reflections harvested from three sets of 20 frames. These initial sets of frames were oriented such that orthogonal wedges of reciprocal space were surveyed. The data collection was carried out using MoKα radiation (graphite monochromator). Randomly oriented regions of reciprocal space were surveyed to the extent of one sphere and to a resolution of 0.84 Å. The intensity data were corrected for absorption and decay using SADABS.³⁶ Final cell constants were calculated after integration with SAINT.³⁷ The structures were solved and refined using SHELXL-97.³⁸ The space groups P2₁/c, P2₁/n and P-1 were determined based on systematic absences and intensity statistics. Direct-methods solutions were calculated which provided most non-hydrogen atoms from the E-map. Full-matrix least squares/difference Fourier cycles were performed which located the remaining non-hydrogen atoms. All non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were placed in ideal positions and refined as riding atoms with relative isotropic displacement parameters. The SQUEEZE function of the program PLATON³⁹ was used to remove 143 effective electrons in diffuse scattering from a volume of 1555.2 ų per cell of 1•O₂P(OPh)₂(BPh₄)₂•MeCN and 120 effective electrons in diffuse scattering from a volume of 1097.4 ų per cell of (1•O₂CC₆H₂-3,4,5-(OMe)₃)₂(BPh₄)₂(OTf)₂•2MeCN. Brief crystal data and intensity collection parameters for the crystalline complexes are shown in Table 1.

Table 1.

Crystal data and structure refinement for 1•O₂PMe₂(BPh₄)(OTf)•MeCN, 1•O₂P(OPh)₂(BPh₄)₂•MeCN, and (1•O₂CC₆H₂-3,4,5-(OMe)₃)₂(BPh₄)₂(OTf)₂•2MeCN

	[1•O2PMe2]-(BPh4)(OTf)•MeCN	[1•O2P(OPh)2]-(BPh4)2•MeCN	[1•O2CC6H2-3,4,5-(OMe)3]2(BPh4)2(OTf)2-•2MeCN
empirical formula	C72H78BF3Fe2N11O6PS	C105H102B2Fe2N11O5P	C160H166B2F6Fe4N22O18S2
fw	1435.99	1762.27	3108.29
T (K)	173(2)	173(2)	173(2)
Mo Kα λ, Å	0.71073	0.71073	0.71073
space group	P21/c	P21/n	P -1
a (Å)	16.0709(9)	23.063(4)	12.576(4)
b (Å)	15.6360(9)	18.942(4)	19.153(7)
c (Å)	28.4703(15)	23.888(4)	36.053(12)
α (deg)	90	90	96.743(5)
β (deg)	103.5800(10)	99.205(3)	92.147(5)
γ (deg)	90	90	103.100(5)
V (Å3)	6954.1(7)	10301(3)	8381(5)
Z	4	4	2
ρ (calc), Mg/m3	1.372	1.136	1.232
abs coeff (mm−1)	0.539	0.352	0.437
R1a	0.0398	0.0446	0.0713
wR2b	0.0984	0.1271	0.1955

^aR1 = Σ||F_o|−|F_c||/Σ|F_o|.

^bwR2 = [Σ[w(F_o²−F_c²)2]/Σ [w(F_o²)²]]^1/2.

Results

X-ray Crystallography

For this work, we synthesized diiron(II) complexes using a variety of O₂X ligands, namely O₂PMe₂, O₂P(OPh)₂, and several aliphatic and aromatic carboxylates, to augment a list that includes O₂AsMe₂, O₂PPh₂, and benzoate. We were able to obtain crystal structures of three of these 1•O₂X compounds (O₂X = O₂PMe₂, O₂P(OPh)₂ and O₂CC₆H₂-3,4,5-(OMe)₃) (Table 2). Like those of previously published complexes,²⁸ the new complexes possess two distorted trigonal bipyramidal iron(II) centers, with N-EtHPTB amine nitrogen atoms and oxygen atoms from the bridging O₂X moiety occupying the axial positions. The equatorial sites are occupied by two benzimidazole nitrogen atoms and the alkoxide oxygen atom (Figure 1). With few exceptions, the respective atoms in the first coordination sphere of each complex reflect approximately equal interatomic distances and angles. Differences of note include average τ values,⁴⁰ O₂X bite distances, O-X-O angles and inter-iron distances. The τ values can be grouped into two sets with τ_ave ~ 0.8 (O₂X = O₂AsMe₂, O₂PPh₂ and O₂PMe₂) and τ_ave ~ 0.9 (O₂X = O₂P(OPh)₂, O₂CC₆H₂-3,4,5-(OMe)₃ and O₂CPh). Comparing bite distances of all complexes, we see O···O distances of ~2.55 Å (X = P) and ~2.23 Å (X = C). The lone X = As complex is in a class of its own with a bite distance of ~2.80 Å. It is also apparent that the O-X-O angle varies with the identity of the X atom (X = As, 113.26 deg.; X = P, 115.17 to 119.47 deg.; X = C, 123.4 to 124.2 deg.). The O-X-O angle measurement correlates inversely with O₂X bite distance; the species with the longest O···O distance (1•O₂AsMe₂) has the most acute O-X-O angle, while the X = C complexes have the shortest bite distances and largest O-X-O angles, with the X = P compounds falling in between these extremes. This inverse correlation reflects variations in X-O bond lengths (As-O_ave = 1.676 Å; P-O_ave = 1.501 Å; C-O_ave = 1.261 Å. Even though the O-X-O angle decreases, concomitant lengthening of the X-O bonds results in a greater O···O distance.

Table 2.

Selected interatom distances and bond angles for [Fe₂(N-EtHPTB)(O₂X)]²⁺.

	1•O2AsMe2a	1•O2PPh2a	1•O2PMe2	1•O2P(OPh)2	1•O2CC6H2-3,4,5-(OMe)3b	1•O2CPha
τave	0.81	0.77	0.82	0.91	0.89	0.93
Interatom distances (Å)
Fe1-O1	2.0145(13)	2.004(2)	2.0184(18)	2.0057(16)	1.967(3)	1.976(5)
Fe2-O1	2.0037(14)	1.992(2)	1.9995(18)	2.0031(15)	1.970(3)	1.964(5)
Fe1-N1	2.3366(16)	2.294(3)	2.323(2)	2.3141(18)	2.328(4)	2.316(6)
Fe2-N2	2.3892(16)	2.352(3)	2.372(2)	2.309(2)	2.314(4)	2.280(7)
Fe1-O2	1.9825(14)	2.008(2)	2.0037(19)	2.0562(16)	2.036(3)	2.057(5)
Fe2-O3	1.9863(14)	2.021(2)	2.0123(18)	2.0506(16)	2.025(3)	2.019(6)
Fe1-N3	2.1372(17)	2.086(3)	2.105(2)	2.0668(19)	2.045(4)	2.064(6)
Fe1-N5	2.1023(18)	2.115(3)	2.127(2)	2.0732(19)	2.046(4)	2.069(6)
Fe2-N7	2.1088(17)	2.075(3)	2.098(2)	2.0551(19)	2.063(4)	2.080(6)
Fe2-N9	2.0759(17)	2.098(3)	2.080(2)	2.075(2)	2.044(4)	2.064(6)
As1/P1/C44-O2	1.6740(15)	1.514(2)	1.509(2)	1.4762(17)	1.254(5)	1.264(9)
As1/P1/C44-O3	1.6776(14)	1.512(2)	1.517(2)	1.4784(18)	1.273(5)	1.253(9)
Fe1···Fe2	3.5357(5)	3.5405(10)	3.5364(6)	3.6211(7)	3.4879(13)	3.4749(31)
O2···O3	2.7991(21)	2.5600(32)	2.5545(28)	2.5520(24)	2.2253(46)	2.2251(74)
Bond angles (degrees)
Fe1-O1-Fe2	123.27(6)	124.76(11)	123.32(9)	129.19(8)	124.73(15)	123.8(2)
O1-Fe1-O2	105.77(6)	105.55(9)	102.50(8)	98.99(6)	98.27(13)	98.6(2)
O1-Fe2-O3	107.92(6)	99.44(9)	103.98(8)	98.93(7)	100.51(13)	101.5(2)
Fe1-O2-As1/P1/C44	127.24(8)	132.01(14)	131.47(12)	133.92(11)	138.1(3)	136.4(5)
Fe2-O3-As1/P1/C44	128.15(8)	138.88(15)	132.54(12)	133.24(10)	134.7(3)	133.9(5)
O2-As1/P1/C44-O3	113.26(7)	115.59(13)	115.17(12)	119.47(10)	123.4(4)	124.2(7)

^aData from reference 28.

^bThe unit cell of this compound contains virtual twins of 1•O₂CC₆H₂-3,4,5-(OMe)₃ produced through a pseudo-inversion center. As there is no substantive difference between the two cations, values listed here are taken from a single cation.

Figure 1.

Crystal structure of the cation 1•O₂CC₆H₂-3,4,5-(OMe)₃ (50% ellipsoids) with hydrogen atoms removed. Generic cartoon of 1•O₂X: O₂X = O₂AsMe₂ (1•O₂AsMe₂), O₂PPh₂ (1•O₂PPh₂), O₂PMe₂ (1•O₂PMe₂), O₂P(OPh)₂ (1•O₂P(OPh)₂), O₂CC₆H₂-3,4,5-(OMe)₃ (1•O₂CC₆H₂-3,4,5-(OMe)₃), O₂CPh (1•O₂CPh).

In four of the six diferrous complexes crystallographically characterized, increased bite distance correlates to increased Fe···Fe distance. With the other two complexes, we observed that: (i) 1•O₂AsMe₂ has an O···O distance ~0.24 Å longer than any other complex, although its interiron distance is within 0.01 Å of those found in both 1•O₂PPh₂ and 1•O₂PMe₂. (ii) 1•O₂P(OPh)₂ has an O···O distance on par with 1•O₂PPh₂ and 1•O₂PMe₂, yet its interiron distance (3.6211 Å) is the greatest of all the compounds, exceeding the distances found in 1•O₂AsMe₂ (3.5357 Å) and the two other X = P species (3.5405 and 3.5364 Å) by ~0.08 Å. This is greater than the ~0.06 Å difference observed between the Fe···Fe distances of the X = C complexes (3.4879 and 3.4749 Å) and those with Fe···Fe distances of ~3.54 Å. The unusually long Fe···Fe distance in 1•O₂P(OPh)₂ is most likely not due to packing effects, because a similar interiron distance of 3.649 Å was reported when 1•O₂P(OPh)₂ crystallized in a different space group (P -1) with different anions (ClO₄).⁴¹

There are two crystallographically interesting observations regarding the compounds we report here and those examined in our previous work.²⁸ First, we note that [Fe₂(N-EtHPTB)( O₂PMe₂)](BPh₄)(OTf)(MeCN) and [Fe₂(N-EtHPTB)(O₂AsMe₂)](BPh₄)(OTf)(MeCN) are crystallographically isostructural. Both fall into the space group P2₁/c with Z = 4 and share virtually identical unit cells with respective values for a, b and c of 16.0709(9) vs. 16.1151(12), 15.6360(9) vs. 15.7370(12) and 28.4703(15) vs. 28.473(2) Å. The α and γ angles are exactly 90 degrees and the β angles are respectively 103.5800(10) and 103.4900(10) degrees. The respective cations, anions and solvent molecules in each unit cell lie in essentially the same positions with the same orientations. The minor variations in unit cell parameters and atom positions arise from differences around the X atom of the O₂X bridging moieties. Small angle changes around X as well as X-C and X-O bond length variations produce minor differences in the cations which slightly alter the position of every other atom in the unit cell due to packing effects. The second observation is that the unit cell containing the cation [Fe₂(N-EtHPTB)( O₂CC₆H₂-3,4,5-(OMe)₃)]²⁺ contains a pseudo-inversion center lying between cations (Figure 2). This effectively doubles the number of unique elements per unit cell, thus doubling the size of the unit cell. The cation “inversion” is reasonably true, but the anions are farther away from the pseudo-inversion center, resulting in greater distortion during “inversion”. Because the differences between the cations are not significant, we chose to focus on one cation rather than redundantly discussing what essentially amounts to duplicate cations.

Figure 2.

ORTEP diagram (50% ellipsoids) of the unit cell containing two molecules of [Fe₂(N-EtHPTB)( O₂CC₆H₂-3,4,5-(OMe)₃)]²⁺ and accompanying anions and solvent molecules (hydrogen atoms removed for clarity). There is a pseudo-inversion center (ic) between the cations. Comparing the cations, we find that the “inversion” is almost true, whereas comparing the BPh₄ ions reveals differences in thermal ellipsoid size and minor differences in atom locations, indicating that the inversion center is not real.

Observation of Two Peroxo Intermediates

Upon reaction with O₂, the virtually colorless solutions of 1•O₂X produce blue-green O₂ adducts (2•O₂X), which in all cases except two convert to deep blue species (3•O₂X) before decaying to yellow final products (4•O₂X). Previously, we reported the formation of 2•O₂PPh₂ and its conversion to 3•O₂PPh₂ at −40°.²⁸ Corresponding studies with O₂X = benzoate afforded only evidence for the formation of 3•O₂CPh under these conditions.^29,42,43 However, visible spectral evidence was obtained for the formation of both 2•O₂CR and 3•O₂CR by going to −90° in CH₂Cl₂, as shown in Figure 3 for O₂X = O₂CCPh₃ with respective λ_max values at 708 and 630 nm. In general, the visible chromophores of 2•O₂X and 3•O₂X fall in the ranges of 630–710 and 580–620 nm, respectively (Table 3), which have been previously assigned to peroxo-to-iron( III) charge transfer transitions for O₂X = benzoate, Ph₂PO₂, and Me₂AsO₂ by resonance Raman spectroscopy.^28,42

Figure 3.

Selected UV-Vis spectra obtained in the reaction of O₂ and 1•O₂CCPh₃ in CH₂Cl₂ at −90° showing initial formation of 2•O₂CCPh₃ (solid green line) and subsequent conversion to 3•O₂CCPh₃ (solid blue line). Dotted lines correspond to spectra leading to the appearance of 2•O₂CCPh₃ within the first 40 seconds, while dashed lines correspond to spectra associated with the conversion of 2•O₂CCPh₃ to 3•O₂CCPh₃ over the course of the following hour.

Table 3.

UV-Vis properties of 2•O₂X, 3•O₂X and 3′•O₂X in CH₂Cl₂.

O2X	λmax (nm), [ε] (M−1 cm−1)
	2•O2X	3•O2X	3′•O2X
O2AsMe2	632 [2100]	-- a	-- b
O2PMe2	674 [1800]	-- a	616 [2400]
O2PPh2	678 [2100]	621 [1800]	592 [2600]
O2P(OPh)2	680 [2800]	576 [3100]	594 [3600]
O2CCPh3	708 [2500]	630 [2200]	595 [3300]
O2CCMe3	706 [--] c	580 [3100]	577 [3800]
O2CPh	704 [--]c	588 [1500]d	592 [2300]
O2CC6H2-3,4,5-(OMe)3	707 [2600]	590 [3200]	592 [4300]
O2CC6H3-3,4-(OMe)2	704 [--]c	578 [6200]	590 [5300]
O2CC6H3-3,5-(OMe)2	706 [--]c	588 [5400]	592 [5600]
O2CC6H4-4-OMe	705 [--]c	580 [3000]	590 [4000]

^aConversion to 3•O₂X did not take place at any temperature.

^bConversion to 3′•O₂X did not take place at any temperature even with addition of 100 equivalents of OPPh₃.

^cConversion of 2•O₂X to 3•O₂X began before complete formation of 2•O₂X, so ε was not determined.

^dValues from reference 42.

Resonance Raman spectra for 2•O₂CCPh₃, 3•O₂CCPh₃, 2•O₂P(OPh)₂, and 3•O₂P(OPh)₂ shown in Figure 4 are representative of the species observed in this study. ¹⁶O₂ and ¹⁸O₂ labeling experiments reveal O-O stretching frequencies (Table 4) that by comparison to those previously published^28,42 allow us to sort them into two clusters. The 2•O₂X cluster exhibits ν_O-O values ranging from 839 to 851 cm⁻¹ that downshift to 791–807 cm⁻¹ upon ¹⁸O substitution, while the 3•O₂X cluster has ν_O-O values ranging from 897 to ~910 cm⁻¹ that downshift to 845–857 cm⁻¹ with ¹⁸O incorporation. On the other hand, no systematic difference is observed for the ν_Fe-O values of the 2•O₂X and 3•O₂X complexes, all of which range from 457 to 479 cm⁻¹ (Table 4). These features can be assigned to the symmetric Fe-O stretch of the Fe–O₂-Fe moiety and often appear as Fermi doublets. The difference in the ν_O-O values of 2•O₂X and 3•O₂X complexes has previously been attributed to a change in the Fe–Fe distance in the O₂ adducts on the basis of a systematic study of 1,2-peroxo-bridged diiron(III) complexes by Fiedler et al.²⁷ For the 2•O₂X and 3•O₂X complexes, the change in the Fe-Fe distance indicated by the difference in the ν_O-O values is postulated to derive from the conversion of the O₂X moiety from a bidentate bridging ligand in 2•O₂X to a terminal monodentate ligand in 3•O₂X (Scheme 1).

Figure 4.

Resonance Raman spectra of 2•O₂CCPh₃ (A), 3•O₂CCPh₃ (B), 2•O₂P(OPh)₂ (C) and 3•O₂P(OPh)₂ (D). Solid red lines (¹⁶O₂) and dotted blue lines (¹⁸O₂); T = 77 K, λ_ex = 647 nm.

Table 4.

Fe–O and O–O stretching frequencies of 2•O₂X and 3•O₂X complexes determined using resonance Raman spectroscopy.

O2X	2•O2X		3•O2X
	νFe-O (cm−1) [18O]	νO-O (cm−1) [18O2]	νFe-O (cm−1) [18O]	νO-O (cm−1) [18O2]
O2AsMe2	464 [433]	845 [796]	--	--
O2PMe2	467, 479 [449, 460]	839 [791]	--	--
O2PPh2	465, 476 [455]	845, 853 [807]	477 [458]	897 [848]
O2P(OPh)2	457, 469 [436, 446]	851 [806]	479 [460]	897 [845]
O2CCPh3	466, 474 [447, 454]	841 [792]	466, 479 [448, 456]	903, 917 [857]
O2CCMe3	--	--	466, 475 [447, 457]	897, 912 [846]
O2CPha	--	--	476 [460]	900 [850]
O2CC6H2-3,4,5-(OMe)3	--	--	465, 476 [449, 457]	900, 914 [853]
O2CC6H3-3,4-(OMe)2	--	--	466, 474 [449, 456]	898, 912 [847]
O2CC6H3-3,5-(OMe)2	--	--	465, 476 [449, 458]	899, 913 [848]
O2CC6H4-4-OMe	--	--	466, 474 [448, 457]	899, 913 [847]

^aValues from reference 42.

Scheme 1.

Conversion of 2•O₂X to 3•O₂X or 3′•O₂X.

Attempts to obtain resonance Raman spectra for either 2•O₂CC₆H₂-3,4,5-(OMe)₃ or 2•O₂CC₆H₃-3,4-(OMe)₂ were unsuccessful, even though they exhibited a longer half life than 2•O₂CCPh₃. A color change was observed during the freezing process. Their respective Raman spectra were comparable to those of 3•O₂CC₆H₂-3,4,5-(OMe)₃ and 3•O₂CC₆H₃-3,4-(OMe)₂, indicating conversion took place during the phase change (Figure S1). On the other hand, the longer lifetimes of the 3•O₂X intermediates allowed us to collect spectra of every one (Figures 4 and S1), with the exception of 3•O₂PMe₂, which did not form in quantities sufficient for characterization. The resonance Raman spectrum of 3′•O₂PPh₂, formed by treatment of 2•O₂PPh₂ with an excess of OPPh₃ (Figure 5) was also acquired. As it was nearly identical to the previously reported spectrum of 3′•O₂CPh,²⁷ we did not carry out the corresponding experiments with other 3′•O₂X intermediates.

Figure 5.

Resonance Raman spectra of 3′•O₂PPh² (¹⁶O₂ = solid red line, ¹⁸O₂ = dotted blue line).

Conversion of 2•O₂X to 3•O₂X

The spectral differences between 2•O₂X and 3•O₂X make UV-Vis absorption spectroscopy an excellent tool for following the transition from the first peroxo intermediate to the second. Based on the data shown in Figure 3, analysis of the spectral changes observed upon oxygenation of 1•O₂CCPh₃ at −90° afforded a first-order rate constant of 1.7 × 10⁻³ s⁻¹ for the conversion of 2•O₂CCPh₃ to 3•O₂CCPh₃. This value is comparable to that previously reported for the conversion of 2•O₂PPh₂ to 3•O₂PPh₂ at −40°, thus implicating a more facile conversion for the carboxylate-bridged species. Figure 6 provides evidence for the formation of two peroxo species at −90° as well even for the well studied complex 1•O₂CPh,^29,42,44 although they are not as spectroscopically distinct as those of 2•O₂CCPh₃ and 3•O₂CCPh₃. Stopped-flow kinetic analysis at −80°C revealed the conversion from 2•O₂CPh to 3•O₂CPh with a k_obs of 5.3 × 10⁻² s⁻¹, much faster than that observed for 2•O₂CCPh₃ to 3•O₂CCPh₃ at −90°. This comparison provides a rationale for why 2•O₂CPh was not observed previously.^29,42,44 The rate of conversion of 2•O₂X to 3•O₂X decreased as electron donating substituents were introduced onto the benzoate ring or as the steric bulk of the carboxylate bridge was increased (Table 5). The rate of conversion was further decreased by replacing the trigonal carboxylate moiety with tetrahedral anions such as (PhO)₂PO₂⁻, Ph₂PO₂⁻, Me₂PO₂⁻, and Me₂AsO₂⁻ (Table 5).

Figure 6.

Selected UV-Vis spectra obtained in the reaction of O₂ with 1•O₂CPh at −90° showing evidence for the formation of two peroxo intermediates. The solid green line represents the spectrum with the largest fraction of 2•O₂CPh formed, while the solid blue line corresponds to the subsequently formed 3•O₂CPh. Dotted lines correspond to spectra leading to the maximum amount of 2•O₂CPh formed within the first 90 seconds after oxygenation, while dashed lines correspond to spectra associated with the conversion of 2•O₂CPh to 3•O₂CPh over the course of the following 30 minutes.

Table 5.

pK_a values of HO₂X and first order k_obs values for conversion of 2•O₂X to 3•O₂X and 2•O₂X to 3′•′ O₂X.

O2X	pKa of HO2X	2 → 3 a Kobs (s−1)	2 → 3′b Kobs (s−1)
O2AsMe2	6.27	-- c	-- d
O2PMe2	3.08 e	-- c	1.8(4) × 10−4 (−40 °C)
O2PPh2	2.32 e	3.1 × 10−4 f (−40 °C)	4.3(4) × 10−3 (−40 °C)
O2P(OPh)2	1.85 e	1.2(1) × 10−4	3.1(1) × 10−2
O2CC6H2-3,4,5-(OMe)3	4.24	4.2(6) × 10−4	0.13(5)
O2CC6H3-3,4-(OMe)2	4.36	6.6(9) × 10−4	0.12(1)
O2CCPh3	3.96	1.7(2) × 10−3	2.6(3) × 10−2
O2CC6H3-3,5-(OMe)2	3.97	4.8(2) × 10−3	0.20(1)
O2CC6H4-4-OMe	4.50	4.7(4) × 10−3	0.12(1)
O2CCMe3	5.03	1.8(1) × 10−2	8.0(3) × 10−2
O2CPh	4.19	5.3(1) × 10−2 g (−80 °C)	-- h

^aAll rates measured at −90 °C except where noted.

^bAll rates measured after addition of 20 equivalents of OPPh₃ at −90 °C except where noted.

^cConversion to 3•O₂X did not take place at any temperature.

^dConversion to 3′•O₂X did not take place at any temperature even with addition of 100 equivalents of OPPh₃.

^eMeasured in 7% EtOH (Reference 53).

^ft −40 °C in CH₂Cl₂, 3•O₂PPh₂ starts to decay before complete conversion from 2•O₂PPh₂ occurs. For this reason, k_obs was calculated from the y-intercept of the OPPh₃ concentration dependence plot for the conversion of 2•O₂PPh₂ to 3′•O₂PPh₂ (Figure S2).

^gRate measured at −80 °C using stopped-flow techniques.

^hBy the time enough 2•O₂CPh had formed to allow for addition of OPPh₃, significant conversion to 3•O₂CPh had occurred, preventing accurate rate determination for the conversion of 2•O₂CPh to 3′•O₂CPh.

In our previous publication,²⁸ we postulated that the conversion of 2•O₂X to 3•O₂X in acetonitrile involves a change in O₂X binding mode from bridging to terminal and coordination of a solvent molecule in position L (Scheme 1). For this work we added OPPh₃ to solutions of 2•O₂X in hopes of substituting it into the L position. Upon OPPh₃ addition, purple-blue species (3′•O₂X) indeed form with λ_max values at wavelengths near each respective 3•O₂X, although the extinction coefficients of the new intermediates are higher than those observed for each respective 3•O₂X (Table 3) with the exception of the O₂CC₆H₃-3,4-(OMe)₂-based species. While facile conversion to 3′•O₂X from most 2•O₂X intermediates was accomplished by adding 20 equivalents of OPPh₃ at −90°, 2•O₂PMe₂ and 2•O₂PPh₂ had to be warmed to −40° to attain reasonable rates of conversion. In contrast, 2•O₂AsMe₂ appears to be unaffected by addition of up to 100 equivalents of OPPh₃, even at temperatures as high as 20°.

After finding evidence indicating the formation of 2•O₂CPh, we opted to further examine the oxygenation kinetics of both 1•O₂PPh₂ and 1•O₂CPh using stopped-flow techniques. 1•O₂PPh₂ offers an excellent opportunity to characterize the initial oxygen coordination with the formation of 2•O₂PPh₂, since the subsequent rearrangement of 2•O₂PPh₂ to 3•O₂PPh₂ is relatively slow. The reaction between 1•O₂PPh₂ and O₂ in either MeCN or CH₂Cl₂, which is accompanied by rapid growth of the absorption band with λ_max ≈ 680 nm, was studied. At low temperatures (down to −80 °C in CH₂Cl₂), the initially formed diiron-O₂ intermediate 2•O₂PPh₂ is stable for at least an hour. However, noticeable decay of this intermediate was observed under intense illumination with polychromatic, UV-rich light of the arc lamp. In order to avoid complications from undesirable photodecomposition, quantitative kinetic measurements were performed in a single-wavelength mode, with low-intensity monochromatic light and a sensitive photomultiplier detector. Under these conditions, photobleaching was not observed. In order to determine the rate law, kinetic traces were acquired under large excess of O₂, showing single-exponential growth of 2•O₂PPh₂ and yielding the values of observed pseudo-first-order rate constants, k_obs. The observed rate constants did not depend on the initial concentration of 1•O₂PPh₂, in agreement with a rate law that is first-order in the diiron precursor. The observed rate constants increased linearly with an increase in the concentration of O₂ (Figure S3), indicating that the reaction was first order in dioxygen. It can be concluded that the formation of 2•O₂PPh₂ is a second-order process (first-order in diiron complex and first-order in O₂): v = k₂[1•O₂PPh₂][O₂]

As temperature was raised, the oxygenation rate of 1•O₂PPh₂ increased modestly, and the conversion of 2•O₂PPh₂ to 3•O₂PPh₂ became pronounced. The Eyring plot (Figure 7) for the oxygen binding step was linear over a broad temperature range (from −80 to −20° in CH₂Cl₂), yielding the activation parameters that are summarized in Table 6. Kinetics of oxygen binding to 1•O₂PPh₂ in MeCN were also examined, and very similar activation parameters were extracted (Table 6). Subsequent disappearance of 2•O₂PPh₂ was readily observed in MeCN at temperatures above −40 °C (Figure S4). The spectral changes agreed well with the conversion of 2•O₂PPh₂ into 3•O₂PPh₂ described in detail by Frisch et al.²⁸ Low activation enthalpies and large negative activation entropies for the oxygenation of 1•O₂PPh₂ leading to 2•O₂PPh₂ are typical of associative O₂ coordination at vacant or labile iron sites,^34,45 which are present in the starting diiron(II) complex. The rates and activation parameters for the oxygenation of 1•O₂PPh₂ in CH₂Cl₂ or in acetonitrile are very similar to previously published kinetic parameters for the oxygenation of 1•O₂CPh in propionitrile.⁴³ The kinetics of oxygenation of 5-coordinate diiron(II) complexes with a related dinucleating ligand HPTP are also very similar (Table 6).⁴⁶

Figure 7.

Eyring plot for O₂ binding to 1•O₂PPh₂ in CH₂Cl₂.

Table 6.

Kinetic parameters for reactions of diiron(II) complexes with dioxygen

Complex	Solvent	T range (°C)	K (−40 °C) (M−1 s−1)	ΔH‡ (kJ/mol)	ΔS‡ (J/K·mol)	Reference
1•O2PPh2	CH2Cl2	−80 to −20	4.7×102	15.7(4)	−124(6)	This work
1•O2PPh2	MeCN	−40 to 0	2.55×103	13.7(4)	−119(10)	This work
1•O2CPh	EtCN	−75 to −15	1.03(12)×102	15.4(6)	−121(3)	43
[Fe2(HPTP)(O2CPh)]2+	MeCN	−40 to 0	7.3 × 103	15.8(4)	−101(10)	46
[Fe2(HPTP)(O2CPh)]2+	CH2Cl2	−80 to 0	67	16.7(2)	−132(8)	46

A closer examination of dioxygen binding to 1•O₂CPh identified the individual reaction steps of the overall reaction. Stopped-flow observation of the reaction of 1•O₂CPh with dioxygen (2.9 mM) in CH₂Cl₂ at −80° reveals a two-step process (Figure S5), in agreement with sequential formation of intermediates 2•O₂CPh and 3•O₂CPh. At longer wavelengths (greater than ca. 650 nm), the initial absorbance increase takes place within 15 seconds and is followed by decay that is complete within 100 seconds. At 640 nm, the initial absorbance increase was also seen, but after 15 seconds the kinetic trace becomes flat. At shorter wavelengths, a biexponential absorbance increase was seen. Two-exponential fit of the variable-wavelengths data gave the following value of the observed rate constant: k₁ = 0.22 s⁻¹ at −80 °C (76 M⁻¹s⁻¹ under an assumption of a second-order process). This rate constant corresponds to the first reaction step, formation of 2•O₂CPh. Interestingly, it is in excellent agreement with the reported value of the rate constant for the reaction of 1•O₂CPh with O₂ (135 M⁻¹s⁻¹ at −75 C in propionitrile, with the activation parameters of ΔH^‡ = 15.4 kJ/mol ΔS^‡ = −121 J K⁻¹ mol⁻¹).⁴³ In that work, 3•O₂CPh was characterized spectroscopically, and it was assumed that the kinetics of oxygenation corresponded to the formation of 3•O₂CPh. Based on our studies, the kinetic parameters reported by Feig et al.⁴³ correspond to the first oxygenation step, the formation of 2•O₂CPh; the next step, conversion of 2•O₂CPh to 3•O₂CPh, is very rapid at most temperatures.

Discussion

One step in the catalytic activation of dioxygen by biological diiron(II) systems often produces (μ-η¹:η¹-peroxo) diiron(III) moieties. Some of these intermediates exhibit a level of stability that allows for them to be trapped and characterized.^9–16 The same is true for many synthetic (μ-η¹:η¹-peroxo) diiron(III) complexes,^{18,25,27,41,47} some of which are so stable they have been crystallographically characterized.^19–22 In previously published work,²⁸ we used the dinucleating ligand N-EtHPTB and different oxyanions (O₂X) to synthesize a set of three dioxygen binding diiron(II) complexes (1•O₂X). Our efforts revealed that the (μ-η¹:η¹-peroxo) diiron(III) intermediates produced upon oxygenation of the diiron(II) precursors come in two forms, green-blue 2•O₂X and deep blue 3•O₂X (Scheme 1). In the initial form, the oxyanion acts as a three-atom bridge between the iron centers and the pendant benzimidazoles are cis to each other on each iron. In many cases, this intermediate converts to a second form, wherein the O₂X ligand has moved to a terminal position, allowing the pendant benzimidazoles of the N-EtHPTB ligand to rearrange from a cis to a trans disposition on each iron. We found that the stability of 2•O₂X is influenced by the identity of O₂X (O₂X = O₂AsMe₂, O₂PPh₂ and O₂CPh) and concluded that the dominant factor governing 2•O₂X stability is the bite distance (O···O) of the O₂X moiety in 1•O₂X as determined by X-ray crystallography. Anions with greater bite distances are better able to accommodate the >3 Å Fe···Fe distance in 2•O₂X and gave rise to more stable 2•O₂X intermediates; for X = O₂AsMe₂, the 2•O₂X intermediate was so stable that no observable conversion to 3•O₂X was observed prior to decomposition.

For this work, we synthesized several more 1•O₂X complexes to assess additional factors that may affect the stability of 2•O₂X. Specifically, we were interested in examining effects produced by electronic and steric changes in O₂X and how those results relate to effects produced by differences in O₂X bite distances. We also investigated how OPPh₃ could be used to destabilize some 2•O₂X intermediates. In addition, we used low-temperature stopped-flow techniques to demonstrate that the oxygenation of 1•O₂CPh was in fact a two-step process like those we have described for the other 1•O₂X complexes. Indeed, 2•O₂CPh is short-lived even at −90°, so the adduct we originally observed at −40° and assigned to be 2•O₂CPh in 1990^29,42 is in actuality the more stable 3•O₂CPh isomer. Finally, we determined the activation parameters for the oxygenation of 1•O₂PPh₂ in CH₂Cl₂ and in MeCN and found them to be very similar to those reported earlier for 1•O₂CPh in EtCN (Table 6),⁴³ strongly suggesting a common rate determining step for these reactions corresponding to the formation of 2•O₂X.

The focus of the experiments reported in this paper has been to gain further insight into the factors that affect the conversion of 2•O₂X to 3•O₂X, in which irreversible conversion to 3•O₂X is preceded by movement of the O₂X moiety from a bridging to a terminal position (Scheme 1). We postulate that there is a rapid equilibrium between bridging and terminal coordination modes of the O₂X ligand in 2•O₂X and only the isomer with a terminal O₂X ligand can undergo conversion to 3•O₂X during which the benzimidazole arms of the N-EtHPTB ligand shift from a cis relationship to each other to a trans configuration on both iron centers. Thus the initial preequilibrium should be affected by the basicity of O₂X ligands, with the more basic ligand favoring the bridging mode and thereby decreasing the fraction of monodentate O₂X isomer available to undergo conversion to 3•O₂X. On the other hand, as the transformation from 2•O₂X to 3•O₂X entails a significant rearrangement of the coordination spheres about each iron center, the conversion should be slowed down by an increase in the steric bulk of O₂X.

The trend in the stability of the 2•O₂X complexes is best discerned by an examination of the kinetic data for the large subset of complexes with carboxylate bridges. A perusal of Table 5 suggests that, while it may be possible to observe an effect of ligand basicity in comparisons of select pairs (e.g. O₂CPh vs. O₂CC₆H₄-4-OMe and O₂CPh vs. O₂CCMe₃), for the most part steric considerations supersede ligand basicity arguments. For example, the stability of 2•O₂CR is enhanced 30-fold in the series, R = Ph < CMe₃ < CPh₃, commensurate with the increase in steric bulk on the carbon atom adjacent to the carboxylate function. Although the higher pK_a (5.03) of HO₂CCMe₃⁴⁸ relative to benzoic acid (pK_a = 4.19)⁴⁸ could be used to rationalize the threefold longer lifetime of 2•O₂CCMe₃, the 30-fold greater stability of 2•O₂CCPh₃ than 2•O₂CPh cannot be explained by the slightly lower pK_a of HO₂CCPh₃ (3.96)⁴⁹ but can easily be rationalized by the much greater bulk of the triphenylmethyl group relative to phenyl.

Complexes with methoxy-substituted benzoate bridges support the above arguments. The conversion of 2•O₂CC₆H₂-3,4,5-(OMe)₃ to its 3•O₂X form is 100-fold slower that for 2•O₂CPh, despite the fact that the two carboxylic acids have essentially identical pK_a values (4.24 and 4.19, respectively).^48,50 Therefore the considerable difference in the stabilities of 2•O₂CC₆H₂-3,4,5-(OMe)₃ and 2•O₂CPh (Table 5) must arise from steric considerations. Other methoxy-substituted benzoate complexes exhibit intermediate rates of conversion; 2•O₂CC₆H₃-3,4-(OMe)₂ converts just slightly faster than 2•O₂CC₆H₂-3,4,5-(OMe)₃, while 2•O₂CC₆H₃-3,5-(OMe)₂ and 2•O₂CC₆H₄-4-(OMe) are 10-fold faster. From these results, it seems clear that additional steric bulk in the meta and/or para positions of benzoate-based oxyanions is the primary factor that stabilizes the resultant 2•O₂X complexes.

The lifetimes of 2•O₂CR complexes are shortened upon addition of OPPh₃. We previously found that OPPh₃ had a good binding affinity for the parent Fe^III ₂(N-Et-HPTB)-peroxo complex and exerted such a significant stabilizing effect that its presence led to the crystallization and structural characterization of [Fe^III ₃(μ-1,2-O₂)(N-Et-HPTB)(OPPh₃)₂]³⁺.²⁰ Addition of 20 equivalents of OPPh₃ to 2•O₂CR at −90° in fact accelerated the conversion to 3′•O₂CR with rates of 0.03–0.2 s⁻¹, representing a much smaller range of values than for the conversions in the absence of OPPh₃ (Table 5). These results suggest that the binding of OPPh₃ to the diiron center facilitates the ligand rearrangement required to convert from 2•O₂X to 3•O₂X (Scheme 1).

Tetrahedral O₂X bridges also increase the stability of 2•O₂X. In this series, we compared complexes with O₂AsMe₂, O₂PMe₂, O₂PPh₂ and O₂P(OPh)₂ bridges that differ in steric properties and basicity (pK_a = 6.27, 3.08, 2.32 and 1.85, respectively).^51–53 The order of stability in this subset is O₂AsMe₂ > O₂PMe₂ > O₂PPh₂ > O₂P(OPh)₂ (Table 4), following the trend of decreasing pK_a values. Only 2•O₂P(OPh)₂ undergoes conversion to the corresponding 3•O₂X form at −90°, with k_obs = 1.2(1) × 10⁻⁴ s⁻¹, a rate that is slower than observed for five of the six 2•O₂CR complexes at −90° (Table 5). While 2•O₂PPh₂ is indefinitely stable at −90°, it converts to 3•O₂PPh₂ when warmed to −40°, at a rate comparable to that of 2•O₂P(OPh)₂ at −90°. Neither 2•O₂PMe₂ nor 2•O₂AsMe₂ generate an observable 3•O₂X form; instead they appear to decay directly to 4•O₂X. However, 2•O₂PMe₂ can be converted to 3′•O₂PMe₂ at −40° by addition of OPPh₃, whereas 2•O₂AsMe₂ does not convert to 3′•O₂AsMe₂, even when reacted with 100 equivalents of OPPh₃ at room temperature. The increasing stability of 2•O₂X with the bite distance of the oxyanion bridge (X = C, 2.23 Å; X = P, 2.55 Å; X = As, 2.80 Å) reflects the ability of the O₂X bridge to span the Fe•••Fe distance required by the dinucleating HPTB ligand framework without imposing a strain on the six-member ring formed by the three atoms of the oxyanion bridge, the two iron atoms and the alkoxide oxygen on the N-EtHPTB ligand. The O₂AsMe₂ complex, having the longest O•••O bite distance, represents the most stable of the 2•O₂X intermediates.

In this paper, we have investigated the various factors that influence the conversion of 2•O₂X O₂ adducts to corresponding 3•O₂X species. This conversion entails the shift of the O₂X ligand from a bidentate bridging mode to a terminal monodentate mode, providing another example of a diiron complex involved in a mechanistically important “carboxylate shift” first recognized by Lippard and coworkers twenty years ago.³⁰ An analogous carboxylate shift has been suggested by Do et al⁵⁴ to occur upon protonation of O₂ adducts of [Fe^II ₂(N-EtHPTB)(μ-O₂CR)]²⁺ complexes (R = Ph or C₆F₅) at −30°. However, although unequivocal evidence was provided for the protonation of the carboxylate ligand, it is clear from a comparison with the spectroscopic data we have presented in this paper that the O₂ adducts studied by Do et al⁵⁴ must be assigned structures we now associate with 3•O₂CR complexes, which have terminal monodentate carboxylates. On the other hand, the corresponding 2•O₂CR complexes (with bridging carboxylates) are in fact formed upon oxygenation of the diiron(II) precursors (as previously assumed^29,42,54) but can be observed only at −90°, as they readily undergo the “carboxylate shift” at that temperature to isomerize to corresponding 3•O₂CR species. However, the key take-home message of the study by Do et al⁵⁴ is that protonation of a carboxylate ligand on a diiron complex can occur and give rise to a complex with a carboxylic acid ligand. Indeed a carboxylate ligand may serve as a convenient conduit for delivering a proton to a bound peroxide in order to facilitate O-O bond cleavage, for example, in the conversion of intermediate P to Q in the case of sMMO for which a proton is clearly implicated.^13,55